The present invention relates to an image reading device for use in a facsimile, a copying machine and a scanner, which device can read an image data from a specified area of an original document by using an original document reading mechanism such as a line sensor for reading image data from an original image at a specified time-interval based on a reference clock and a mechanism for moving the original document or the image reading mechanism.
Conventional facsimiles and copying machines, which read an original document image by using an image reading system using a line sensor consisting of linearly arranged image sensors such as CCDs, involve such a common problem that a distorted image may be produced due to fluctuation (variation) of document image scanning speed since the document or the reading head is mechanically moved in the feed (vertical scanning) direction.
Japanese Laid-Open Patent Publication No. 4-196964 discloses an image reading device that can detect a change in the scanning speed (fluctuation) by using a flowchart as shown in FIG. 1.
The image reading device reads a document image together with an equal-pitch scale (reference scale) consisting of high-density and low-density lines of the same width, which lines are arranged alternatively at the same distance in the feed direction. Namely, its image-reading means such as a line sensor read the document image and the equal-pitch scale at the same time, detects read-time intervals of respective line-to-line distances of the scale (i.e., each distance from a high-density line to subsequent low-density line) and compares the detected time-intervals with the known reference value to determine the scanning speed fluctuation.
The essence of the art described in Japanese Laid-Open Patent Publication No. 4-196964 consists in that the fluctuation of the scanning speed is determined by reading the image together with an equal-pitch scale disposed in the feed direction and comparing read-time intervals of the scale lines with the known reference value. This method can improve the accuracy of image scanning to compensate the possible distortion of a reproduced image due to fluctuation of the scanning speed without a considerable increase of the manufacturing cost of the image reading device.
However, the line sensor used in the above-described device is driven by clock signals and reads a document image at a specified time-interval in synchronism with the clock signals.
Consequently, a change in the scanning speed may not be detected if the displacement due to the scanning speed change of the scanner is smaller than one pixel in the feed direction. To increase the resolving power of the line sensor for detecting the scanning speed fluctuation, it is necessary to considerably increase the clock frequency. This solution, therefore, is impractical in view of increasing high-frequency noise and additional cost.
In case of reading the reference line scale by using analog elements (photo diodes or photo transistors) in combination with a specialized A/D converter, time must be measured by using a high-speed A/D converter, a high-speed timer and a high-response optical device. This may also arise the problems of increasing high-frequency noise and increasing the cost.
Furthermore, the use of a higher frequency clock cannot always enable the line sensor to finely read the equal-pitch line scale for analysis of the speed fluctuation because the size of an image read by one pixel of the line sensor is determined by an optical lens system interposed between a document and the line sensor. Namely, the line sensor may not detect the scanning speed fluctuation if the displacement due to the fluctuation is smaller than one pixel in the feed direction.
The reference line scale used in the conventional image reading systems is an arrangement of lines alternating black (high-density part) and white (low-density part) with the same width. There may arise a problem that a half-pitch shift of the reading pitch frequency of a scanner in the feed direction from the frequency of the equal-pitch lines of the reference scale may result in outputting a halftone image which cannot be analyzed.
The use of an equal-pitch scale having lines with width not less than two pixels of the line sensor for solving the above problem may make the system unable to detect scanning speed fluctuation.
Analysis of the scanning speed fluctuation by integration of results of obtained by processing an image from its top with a comb filter may have an incorrect result if read image data contains erroneous data due to a small change in the density of the equal-pitch scale image and the presence of dust and noise when reading the image.
In addition, the above calculation premises that the same outputs appear continuously to represent black or white lines of the reference scale image. However, the black and white lines of the scale may have small fluctuation of output values in may case by the effect of adjustments of the line sensor and the A/D converter and the above-mentioned errors.
Therefore, the white line has width equal to two pixels can not always have the same values by two successive pixels at the output of the A/D converter. Speed fluctuation component cannot be separated and extracted if the two white lines of the reference scale have different output values.
The above problem may be avoided by applying a single correction factor to all image data so that all white parts and all black parts may have respective fixed values. However, this solution may cause partial excessive or insufficient compensation if a white shading level or a black shading level is preset to vary while the scanner scans image data in the feed direction. The accuracy of analysis may be decreased.
Furthermore, when the corrected values are not only used for separating a black line part from a white-line part but used for calculating the scanning speed fluctuation, the analysis may not attain the sufficient accuracy depending on the corrected state.
It can be also sought to observe the scanning speed fluctuation by extracting a two-dot part showing the same output values in series and determining a value of this position by extracting an integrated value of the part exiting ahead by two dots from the extracted output value. This method, however, has a final result of subtraction at a portion from a black-line to a white-line or at a portion from a while-line to a black-line. For white or black lines each having a width of two dots or more, final values are spaced with the white or black part width, zeroing data for each spacing.
Consequently, only a part of data obtained by processing outputs of the A/D converter with a comb filter is used and remaining data is discarded. No improvement is achieved even by holding a preceding value for each zero-data part.
In other words, the analysis with a chart having a widened frequency (e.g., black line of 1 dot in width and a white line widened to x dots in width) can effectively use only 2/(x+1) of measurement results.
The object of the present invention is to provide a document-image reading device that can separate and extract a speed fluctuation component by determining a gray-level change of each scale line (not reading time-interval), analyze a state of speed fluctuation of scanning in the feed direction and, based on the analysis result, eliminate the possibility of image deformation that may be caused from the unstable factors of the driving mechanism of the device.
Namely, a speed fluctuation component is separately extracted by reading a change in gray level of respective scale lines and analyzed a speed fluctuation state of a scanner scanning in the feed direction and deformation of an image can be eliminated based on the analysis result.
Another object of the present invention is to provide a document-image reading device that can separately extract speed fluctuation of its scanner when working with an equal-pitch scale representing alternations of a black line of one pixel in width and a white line of two pixels in width.
Another object of the present invention is to provide a document-image reading device that can perform high-accuracy high-resolution analysis of the scanning speed fluctuation with no affection of partial abnormal data to an entire analysis result.
Another object of the present invention is to provide a document-image reading device that can accurately analyze the speed fluctuation without decreasing the accuracy due to dispersion of reading a scale image.
Another object of the present invention is to provide a document-image reading device that can accurately analyze the speed fluctuation without decreasing the accuracy in the case when a shading level varies during scanning operation.
Another object of the present invention is to provide a document-image reading device that can always perform accurate correction of image data, eliminating the possibility of decreasing the accuracy of analysis depending the correction state.
A document-image reading device of the present invention comprises document-image reading means for reading a document image data at a specified interval given by a reference clock by moving the document-image reading means or the document in the feed direction, which means is further provided with a reference chart reading means for reading a reference chart portion representing alternations of a high-density division and a low-density division arranged in the feed direction at an interval equal to a least pitch readable by the document-image reading means and processing means for processing image data read from the reference chart by using a comb filter followed by integration.
The construction of the device can attain higher accuracy of the speed fluctuation analysis in comparison with a method for measuring time-intervals between lines of an image read from a reference chart (equal-pitch scale) portion.
A document-image reading device of the present invention comprises document-image reading means for reading a document image data at a specified interval given by a reference clock by moving the document-image reading means or the document in the feed direction and is featured in that the document-image reading means is further provided with a reference chart reading means for reading a reference chart portion representing alternations of a high-density division and low-density division, in which every high-density or low-density divisions has a width equal to a least readable pitch and every low-density division has a width equal to an integer-multiple of the least readable pitch, thus the high-density divisions and low-density divisions alternately arranged one after another at respective equal intervals in the feed direction, and processing means processing the image data read from the reference chart by using a comb filter followed by integration, extracting continuous detection of the substantially same integrated values by the integer number of times and determining an integrated value at the extracted position by reducing the extracted integrated value by an integrated value at an upstream point at a distance of the integer-multiple of the least-readable pitch along the feed direction.
This device can therefore eliminate the possibility of making the image data of the equal-pitch scale become an unanalysable 50% gray scale when the reading pitch of the CCD in the feed direction deviates from the pitch of the equal-pitch scale by a 1/2 dot.
Furthermore, the device can obtain higher accuracy of analysis because the analysis can be made by subtracting only peripheral portions when integrated values are the same subsequent values. Namely, the affection of an abnormal value of the partial image on a whole analysis result can be surely avoided.
A document-image reading device of the present invention comprises a document-image reading means for reading image data from a document at a specified time-interval based on a reference clock by moving said document image reading means or the document in a feed scanning direction and is featured in that the document-image reading means is further provided with reference-chart reading means for reading a reference chart representing alternations of a low-density or high-density divisions with a width equal to a least readable pitch and a high-density or low-density division with a width equal to an integer-multiple of the least readable width along the feed direction and image-data processing means for separating the high-density divisions and low-density divisions from the read image-data, determining a correction factor by which image data for all low-density or high-density divisions or respective low-density or high-density divisions can be corrected to have a constant value, multiplying the image data by the correction factor, processing the image data read from the reference chart by using a comb filter, performing integral calculations on the filtered data, extracting continuous detection of the substantially same integrated value by the integer number of times and determining an integrated value at the extracted position by reducing an integrated value, said value extracted from uncorrected image data by an integrated value of an upstream point at a distance of the integer-multiple of the least-readable pitch along the feed direction.
This device can therefore analyze the speed fluctuation of the document-image reading means (e.g., a scanning device) even when the white levels of the equal-pitch scale image cannot indicate a constant value due to insufficient adjustment of a signal processing system for an image reading section of the scanner. Furthermore, its calculation load can be lessened since only one correcting value is used for correcting the white levels of the whole image.
The device can also examine the white levels of the equal-pitch scale image data and determine a specified correction value separately for each white portion, by which the portion may be corrected to have a preset constant value. This makes it possible to analyze the speed fluctuation without decreasing the analysis accuracy even if the white level of the scanner varies during a scan.
In this instance, the calculation load increases but the white level is real time measured and a correction value can be immediately determined with no need of waiting until the whole image data are measured. The speed fluctuation analysis result can be, therefore, outputted just after reading the image.
A document-image reading device of the present invention is further featured in that the image-data processing means is provided with processing means for determining a correction factor, multiplying image data by the correction factor, processing the corrected image data with a comb filter and by integration, detecting the substantially same values in succession by the integer-number of times, reducing an integrated value, said value extracted from the uncorrected image data, by an integrated value found at a upstream point in the feed direction at a distance of the integer-multiple of the least readable pitch and determining the reduced value as an integrated value of the extracted point.
The device uses corrected data only for determining the conditions of the speed fluctuation analysis and uses uncorrected data for determining speed fluctuation, then determines correction coefficients that make all white (or black) portions have a constant value. It can therefore realize higher accuracy of the speed fluctuation analysis and less amount of calculation. The real time processing is also possible by determining correction values for respective white (or black) portions although the calculation amount increases.
A document-image reading device of the present invention is further featured in that image-data processing means processes image data with a comb filter followed by calculation of absolute values of the filter-processed data and temporarily store the processing result in storing means, and, at the same time, performs processing a read image of a reference chart with a comb filter and by integration, detects the substantially same values in series by the integer-multiple times, reducing an extracted integrated value by an integrated value found at the upstream-side point in the feed direction at a distance of the integer-multiple of the least readable pitch, determines the reduced value as an integrated value of the extracted point by setting 1 for a positive resultant, 0 for a zero resultant and xe2x88x921 for a negative resultant, then multiplies the resultant values by the processed results temporarily stored in the storing means.
The device calculates speed fluctuation by using absolute values of the resultant data obtained by processing the output of A/D converter with a comb filter. Consequently, the finally obtained result of analysis may reflect each part changing from a white portion to a black portion and/or each part changing from a black portion to a white portion and, therefore, can attain increased accuracy.
The processing may also be conducted after correction of white shading for all image data by using the same correction factor. By doing so, the speed fluctuation analysis can be done even if the white levels of the equal-pitch scale did not meet a constant value due to insufficient adjustment of an image-reading signal-processing system of the scanner. Furthermore, the white level of the equal-pitch scale image-data is corrected by using only one correction value for a whole image, thus achieving the saved amount of calculation.
A document-image reading device of the present invention is further featured in that absolute values obtained by processing with a comb filter before shading correction are used for multiplication.
Accordingly, the corrected values are used only for separating white portions from black portions and the uncorrected data is used for calculation of the speed fluctuation. The device can therefore attain high accuracy of analysis.
The device can determine correction factors for correcting respective white line portions to have a full-bit output of an A/D converter and can separately correct the respective portions with corresponding correction factors since the corrected values are used only for separation of white portions from black portions and not-yet-corrected data is used for calculating the speed fluctuation.
Therefore, this device can accurately analyze the speed fluctuation even if the shading level varies during scanning operation, attaining high accuracy of analysis.